Why isn't this sed regex matching?

Question

I've recently started learning sed. I did

$seq 10 | sed '/[^049]/d'

I was expecting

4 
9
10

as output. But I got

4
9

Where am I making mistake in understanding this regex?

cjm · Accepted Answer · 2011-04-27 02:21:15Z

4

The 1 in the number 10 matches [^049] so it's deleted.

cjm

27.2k12 gold badges91 silver badges84 bronze badges

answered Apr 27, 2011 at 2:02

Mark McKinstry

15.4k4 gold badges35 silver badges27 bronze badges

Whoops. That was foolish of me to ask here. :( Thanks by the way.
– Dharmit
Apr 27, 2011 at 2:09
1

Minor tip: Bracketing your regex with ^ and $ is useful to ensure that you're always looking at the whole string.
– l0b0
Apr 27, 2011 at 9:48

Add a comment |

pepoluan · Accepted Answer · 2011-04-27 14:51:16Z

1

If you really want to show lines containing '0', '4', or '9', here's how:

seq 10 | sed -n '/[049]/p'

The -n instructs sed to not print any lines. The p command instructs sed to print lines matching the /regex/

Alternatively, you can always use grep :-)

seq 10 | grep -E "[049]"

answered Apr 27, 2011 at 14:51

pepoluan

1,3332 gold badges16 silver badges26 bronze badges

This is valid, but the exact approach the OP was looking for was probably: seq 10 | sed '/[049]/!d' This deletes any lines that do not match [049] rather than deleting any lines which match "not 049", aka [^049].
– Wildcard
Nov 19, 2015 at 0:20

Add a comment |

2 Answers 2